This MIRA proposal details a research program that centers around the development and application of improved, thermodynamically accurate computer models for simulating RNA 3D structures at atomic resolution. These models differ from existing models for RNA in that they are calibrated to reproduce solution thermodynamic data on the physical behavior of nucleotides and nucleosides, an approach that is readily extended to include the effects of unnatural RNAs and RNA-ligand interactions. This technology is particularly important as many biomedically important RNAs are not amenable to traditional structural biology techniques, which makes it difficult to establish basic structure-function relationships that must be understood before potential therapeutic interventions could be designed. Often, the only available structural information on an RNA of interest are secondary structure estimates from bioinformatics or from SHAPE chemical probing experiments. This proposal builds on recent successes in using molecular simulations restrained by sparse SHAPE or NMR data to simulate the folding pathway of a co-transcriptionally folded RNA, as well as describe how the flexibility of microRNA/mRNA complexes affect how they bind the hAGO2 protein. Building on these recent results, a comprehensive research program is proposed in three major parts. The first is the use of alchemical free-energy calculations to measure the energetics of RNA base-pairing and recalibrate them against experiment. The second is a two-dimensional replica-exchange method for fully automated, adaptive RNA folding incorporating variable strength secondary structure constraints ? a method that show promising results that we expect to scale to large (50-100 nt) RNAs including tertiary motifs. Lastly, we propose a novel multi-dimensional technique to simultaneously fold RNA aptamers while also binding small-molecule ligands using Hamiltonian replica-exchange combined with alchemical free energy calculations ? which will be necessary to capture the ?induced fit? of the RNA aptamer upon ligand binding. These calculations will be used to predict ligand binding modes and engineer optimal RNA biosensors through targeting incorporation of chemically modified nucleic acids.